As the conventional technique for realizing the stable system operation even if the load on a computer system is increased, a load sharing has been developed. Particularly, such load sharing is employed for a server in the computer system. On the other hand, there is a situation where previously inputted data may disappear depending on the processing contents, if a different server is allocated every time of access. Accordingly, there have been developed techniques called a session maintenance and a persistence process in which when the load sharing is made, access is continuously allocated to the same server, using the IP address or cookie (Cookie) of the source.
These technical contents were disclosed in non-patent document, Katsuji Yamashita, “Using the load sharing apparatus (1)”, Nikkei Communication, Nikkei BP Company, 2003 Jun. 9, Vol. 392., p. 126-129.
However, even if the above technique of session maintenance as conventionally known was used, one information processing apparatus and the other information processing apparatus could not be allocated to the same server in the load sharing. Accordingly, when the peer-to-peer communication is established between two information processing apparatuses, for example, there was a problem that it is difficult for the two information processing apparatuses to exchange information between information processing apparatuses using the same server in the system for load sharing.
As a method for solving the problem, the information may be shared among a plurality of servers to which access is allocated by load sharing, in which there is another problem that a large load for sharing information is imposed on each server.
Also, there was a demand for lightening a process for session maintenance for use with the conventional technique for session maintenance.